Biostatistics Flashcards
SEM
Standard Error of the Mean
Calculation: SEM = SD / sqrt (sample size)
In any normal distribution, 95% of the sample means lie within +/- 2 SEM of the population mean
95% CI
Mean +/- 1.96 x SEM
*95% certainty that the true mean of the entire population studied will be within this interval
CI for continuous variables vs OR/RR
If continuous: NS if contains 0;
If OR/RR: NS if contains 1
Nondirectional: Difference Hypothesis Testing
Traditional 2-sided t-test
Confidence intervals
Nondirectional: Equivalence Hypothesis Testing
Two 1-sided t-test (TOST) procedure
Confidence intervals
Directional: Superiority Testing
Traditional 1-sided t-test
Confidence intervals
Directional: Noninferiority Testing
Confidence intervals
Student unpaired t-test
Parametric, 2 independent samples/groups, continuous data
group 1 vs group 2
Student paired t-test
Parametric, 2 paired/matched samples, continuous data
group 1: measurement 1 vs measurement 2
One-way ANOVA
Parametric, 3 or more independent samples/groups, continuous data
(group 1 vs group 2 vs group 3)
Two-way ANOVA
Parametric, 3 or more independent samples/groups + an additional factor added, continuous data
(young groups: group 1 vs group 2 vs group 3
old groups: group 1 vs group 2 vs group 3)
Repeated-measures ANOVA
Parametric, 3 or more matched samples (not independent), continuous data
(Group 1: measurement 1 vs measurement 2 vs measurement 3)
Post-Hoc ANOVA tests
Determine which ANOVA groups actually differ from each other
Turkey HSD, Bonferroni, Scheffé, Newman-Keuls
ANCOVA
Parametric, provides a method to explain influence of a categorical (independent variable) on a continuous variable (dependent variable) while statistically controlling for other variables (confounding)
Wilcoxon rank sum, Mann-Whitney U test, Wilcoxon Mann-Whitney Test
Nonparametric, 2 independent samples (related to t-test), ordinal data
Kruskal-Wallis one-way ANOVA by ranks
Nonparametric, 3 or more independent groups (related to one-way ANOVA), ordinal data
Sign test and Wilcoxon signed-rank test
Nonparametric, 2 matched/paired samples (related to paired t-test), ordinal data
Friedman ANOVA by ranks
Nonparametric, 3 or more matched/paired groups, ordinal data
Chi-square test
2 independent groups, Nominal data
Fisher exact test
2 independent groups, Nominal data
*specialized version of chi-square test for small groups (cells) containing less than 5 predicted observations
McNemar
2 paired/related samples, Nominal data
Mantel-Haenszel
Nominal data, controls for the influence of confounders
Pearson Correlation
Strength of relationship of 2 variables (normally distributed continuous data)
Spearman Rank Correlation
Strength of relationship of 2 variables (ordinal data or nonparametric continuous data)
Simple linear regression
1 continuous independent variable, 1 continuous dependent variable
Multiple linear regression
1 continuous independent variable, 2 or more continuous dependent variables
Simple logistic regression
1 categorical response (dependent) variable and 1 continuous/categorical explanatory (independent) variable
Multiple logistic regression
1 categorical response (dependent) variable and 2 or more continuous/categorical explanatory (independent) variable
Nonlinear regression
variables are not linearly related
Polynomial regression
any number of response and continuous variables with a curvilinear relationship (cubed, squared)
Cochran Q
2 or more paired/matched samples, nominal data